Project Summary We propose to create computational methods and tools to reconstruct bacterial haplotypes directly from next generation sequencing (NGS) reads. There often exist different bacterial haplotypes mutated from the same original haplotype in an infected clinical sample. Because of the presence of different haplotypes, drugs effective to the original haplotype may have no effect on certain mutated haplotypes. Those haplotypes that are resistant to an originally effective drug may escape the scrutiny of the drug treatment so that this drug becomes essentially ineffective for the clinical condition. It is thus imperative to identify the resistant bacterial haplotypes from the mixture of different haplotypes. The proposed research will focus on developing novel computational approaches and tools to reconstruct the haplotypes from NGS reads generated from an infected sample. If successful, the proposed research will save time and cost to reconstruct drug-resistant bacterial haplotypes.